GLOBALIZATION IN THE U.S. AUTO INDUSTRY: INTERNATIONAL MERGERS AND ACQUISITIONS AS DRIVERS OF INFORMATION SHARING

Transcription

1 GLOBALIZATION IN THE U.S. AUTO INDUSTRY: INTERNATIONAL MERGERS AND ACQUISITIONS AS DRIVERS OF INFORMATION SHARING Paul N. Isely 1 Grand Valley State Universy Gerald P. W. Simons 2 Grand Valley State Universy simonsg@gvsu.edu ABSTRACT There is growing interest in the role that corporate restructurings play in innovation. Although existing research has looked at the impacts of leveraged buyouts of companies on R&D spending and profs, few studies have looked at the effect that buyouts have on the output of R&D, and those studies focus on the impacts of domestic M&As. This paper contributes to the lerature by empirically investigating the patenting behavior of the Big 3 U.S. automobile companies and the effects of international knowledge spillovers from their foreign acquisions. A log-log regression specification is used to isolate flows of knowledge for between the country of the acquired subsidiary and the U.S. manufacturer. The Lexis-Nexis patent database is used to count international patent cations and inventor country of origin as measures of international information flows. Results indicate that there is a significant difference for the automobile industry in the information flows arising from M&As in two distinct locales: Germany, and the U.K. and Canada, wh the latter pair giving greater knowledge spillovers for low levels of M&A, whereas the former gives greater spillovers as the volume of M&A activy increases. Key words: M&As, innovation, patents, automobile, spillovers. JEL codes: L60, G34 1 Other contact information: Associate Professor of Economics, Grand Valley State Universy, 401 W. Fulton St., Grand Rapids, MI 49504, Phone: (616) Fax: (616) Other contact information: Associate Professor of Economics, Grand Valley State Universy, 401 W. Fulton St., Grand Rapids, MI 49504, Phone: (616) Fax: (616)

2 I. INTRODUCTION The U.S. automobile industry is a textbook example of an old economy industry that has been struggling under the pressures of globalization. Both Ford and General Motors, once icons of American industry, were downgraded to junk-cred status in mid-2005, and a few months later Delphi became the largest automotive company in U.S. history to declare bankruptcy. Throughout s history, firms in the U.S. auto industry have been active in undertaking mergers and acquisions (M&As) wh both domestic and foreign companies. These auto manufacturers have also aggressively patented their innovations. But have these M&As helped the companies innovative process? As the competive pressure that globalization brings from overseas threatens to decimate the U.S. auto industry, are U.S. automotive manufacturers effective in utilizing information and innovation skills from their foreign counterparts? This paper addresses these pertinent issues. II. LITERATURE REVIEW Corporate restructuring, both internal (such as reorganization of company divisions) and external (such as M&As), is generally used as a method to increase company performance. However, the empirical evidence that such performance enhancements are indeed realized is limed 3. Corporate restructuring is also seen as a key determinant of a company s innovative activy 4. But, though M&As are thought to posively influence innovation, there are few empirical studies dealing explicly wh this relationship. Van Beers and Sadowski (2002) analyze the link between M&As and the probabily of innovating for Dutch firms. They find that firms that acquire and divest assets have a greater probabily to innovate but the effect for firms that only acquire is not significant. They conclude that firms that buy and sell others do so to aid innovation. Similarly, Granstrand and Sjölander (1990) look at Swedish technology firms buying smaller firms to obtain knowledge and find many can only access certain markets by M&As wh those having an established presence there. Other studies indicate that M&As do not reflect a push to enhance innovation rather another motivation is the impetus and any impact on innovation is secondary. Ghosh (2004) finds a major M&A consideration for U.S. firms is increasing market share. As firms tend to increase R&D spending wh market share, this implies that acquisions indirectly enhance innovation. Sundarsanam (2004) argues instead that the primary motivation is revenue 3 For example, Ghosh (2001) finds no evidence that a company s operating performance improves after an acquision. 4 See, for example, Geroski et al. (1993). 2

3 growth. Again, as data show R&D spending increasing wh revenue, this also implies that M&As enhance innovation. Danzon et al. (forthcoming) find mergers in the U.S. pharmaceuticalbiotechnology industry to be due to excess capacy concerns in large firms and financial woes in small firms. They show that mergers cause slower growth in R&D for small firms, but have no significant impact on R&D for large firms. There is scant research on how foreign M&As affect R&D. Belderbos (2001) studies how international M&As by Japanese firms affect their overseas R&D and finds the M&As lead to more foreign patents. This paper differs as the focus is on the impact of foreign M&As on a firm s domestic innovative behavior. Isely and Simons (2002) investigate the impact that international information flows (measured by international trade and foreign patent cations) have on patenting in the U.S. auto industry, and find spillovers to be posive from Germany but negative from Japan. They do not use data on M&As. How innovation is measured varies, from industry surveys (as in Van Beers and Sadowski, 2002) to R&D spending (as in Granstrand and Sjölander, 1990). This paper uses patenting as a measure of innovation and R&D output, and foreign patent cations (and inventor location) as a measure of international information flows. There is substantial research on patents as a measure of innovation and cations as a measure of knowledge spillovers 5. Jaffe et al. (2000) find that aggregate cation flows can be used as proxies for knowledge-spillover intensy (p. 218). Jaffe et al. (1993) find a greater likelihood that U.S. patents will ce other U.S. patents rather than foreign ones and a greater likelihood that these cations and patents will come from the same state and metropolan statistical area. Jaffe and Trajtenberg (1999) find that a firm s patents have a greater likelihood of cing each other and inventors are 30-80% more likely to ce patents from the same country than from others. Wh foreign M&As firms can access differentially localized information sets. Given such spatial aspects, international M&As could lead companies to better utilize information from the acquired firms countries. To summarize the current lerature: (i) there is disagreement as to how important enhanced innovation is in the M&A decision; (ii) there is, however, some indication that a firm s innovation is posively affected by s domestic M&As; and (iii) there is also evidence of spatial relationships in patenting. The contribution of this paper is in a combination of characteristics which differ from the existing research: (i) empirical analysis of M&As and innovation in an old economy industry; (ii) a focus on international M&As and international information flows; and (iii) patent data as a measure of information spillovers. 5 For example, Jaffe [1986], Pavt and Soete [1997], Trajtenberg et al. [1997] and Jaffe et al. (1998). 3

4 The choice of industry may appear odd is frequently newer industries, such as semiconductors or biotech, which are highly competive in innovation. Indeed, recent studies focus on such sectors for those very reasons 6 and Giedeman et al. (2006) show that innovation in high tech industries follows a different pattern than in the auto industry. However, one would expect high tech competive industries to readily utilize knowledge from all available sources, including their overseas acquisions. But does an old economy industry, comprising some of the world s largest companies, harness the innovative potential of foreign acquisions? III. DATA AND METHODOLOGY This study looks at the patenting behavior of the Big 3 U.S. auto companies Chrysler Corp., Ford Motor Co. and General Motors Corp. and tries to capture the effects of knowledge spillovers from their foreign acquisions/subsidiaries. It is assumed that knowledge flows follow a standard production function where the inputs are the firm s research activies and the operations of s foreign subsidiaries. R&D spending proxies for a firm s research inputs, wh data from COMPUSTAT, a company s automobile production in a foreign country proxy for s exposure to overseas subsidiaries, wh data from Ward s Automotive Yearbook ( ). Using such production numbers focuses the study on direct exposure to a particular country (e.g. Germany), while profs or revenues reflect sales whin a particular region (e.g. Europe) and cannot be separated by country wh consistency. The impact of foreign M&As on innovation for each company is measured by the sum of s patents created in the foreign country and s patents that ce that country. The U.S. patent process involves a search of existing patents. Relevant prior art is listed on granted patents in the form of cations along wh s country of origin. Data on the number of U.S. patents assigned to the Big 3 are obtained from Hall et al. (2001) and the Lexis-Nexis patent database. The Lexis-Nexis database is also used to count international patent cations and patents for which the inventor/co-inventor is in a foreign country. Research indicates that information flows more easily between countries wh a common language 7. The foreign subsidiaries of the Big 3 are located in several nations, but their varying concentrations over the time period of this study yield the following language groupings for which the three auto companies have consistent subsidiary exposure: (1) Germany; (2) Australia, Canada, New Zealand and the U.K. However, Australia and New Zealand are dropped because of data inconsistencies. 6 See Hall and Ziedonis (2001), Bessen and Hunt (2004) and Danzon et al. (forthcoming). 7 See Jaffe and Trajtenberg (1999). 4

5 As in Jaffe (1986) a log-log specification is used. The estimation equation is: TOTAL = α + β ( R & D ) + β ( PRODUCTION ) + β ( PATENTS 3 β ( YEAR) + β ( PRODUCTION 5 + β ( FORD 8 1 jt 6 ) + β ( R & D ) + β ( CHRYSLER )( PRODUCTION ijt )( GERMANY jt ) ijt ijt ) + ) + β ( GERMANY 7 (1) ) where the variables are defined as follows 8 : TOTAL is the number of patents granted to firm i in year t that eher originated in the target country or ced the target country R&D is firm i s monetary contribution to the creation of new products in year t, deflated by firm i s total patents. PRODUCTION ijt is the number of automobiles produced by firm i in country j in year t, deflated by firm i s total patents. PATENTS is the total patents granted to firm i in year t (regardless of country of origin or cations) YEAR is a trend variable equal to the calendar year. GERMANY is a dummy variable set to one if the country of production is Germany for firm i in year t. FORD jt is a dummy variable set to one if production in country j in year t is by Ford. CHRYSLER jt is a dummy variable set to one if production in country j in year t is by Chrysler. Subscript i designates firms; j countries; and t years. All monetary amounts are measured in real millions of dollars using the implic price deflator wh a base year of 1996, and non-dummy variables are measured using natural logarhms. The time frame is , chosen for years wh consistent data reporting in COMPUSTAT and Lexis-Nexis. Although final decisions on patent applications are made by the U.S. Patent and Trademark Office whin 2 years on average now, the process can take much longer. The end date of 2000 allows for the inclusion of patents granted in The summary statistics are given in 8 The regression was also run wh a GERMANY*R&D*PRODUCTION variable to fully interact German production. This gave no qualative change in the estimated relative elasticies. 5

6 Table 1. Summary Statistics: Variable Observations Mean Std. Dev. (not logged) TOTAL R&D PRODUCTION PATENTS YEAR GERMANY FORD CHRYSLER GM Table 2. Expected Signs of Coefficients: Coefficient on variable Expected Sign R&D + PRODUCTION + PATENTS + R&D*PRODUCTION + YEAR + PRODUCTION*GERMANY? GERMANY - FORD? CHRYSLER? Table 2 gives expected signs of coefficients reflecting the following rationales: (i) The posive sign on R&D is because increases in R&D spending per patent imply larger patents encompassing more information and thereby providing greater information flows; (ii) the posive sign on PRODUCTION is because greater exposure in a country relative to the number of patents means greater exposure to the knowledge of that work force; (iii) the posive sign on PATENTS is because the greater a company s patenting activy, the more likely uses information from other sources, including those overseas; (iv) the posive sign on R&D*PRODUCTION is because of synergies between exposure to international ideas and investment in research; (v) the posive sign on YEAR is because over time electronic search capabilies have improved leading to more patent cations; (vi) the negative sign on GERMANY is because of the language mismatch wh the home company in the U.S.; (vii) there is insufficient information for a reasonable expectation for the PRODUCTION*GERMANY, FORD and CHRYSLER coefficients. 6

7 IV. EMPIRICAL RESULTS Table 3. Regression Results: Regression Coefficient Intercept (1.65) R&D 2.52 (59.85)** PRODUCTION 0.88 (4.71)* R&D*PRODUCTION (5.21)* PATENTS 1.23 (3.28) + YEAR (1.85) PRODUCTION*GERMANY 0.15 (3.49) + GERMANY (3.58) + FORD 0.43 (3.31) + CHRYSLER 0.26 (0.42) R-square.936 Observations 97 Number of Companies 3 Absolute value of t statistics in parentheses + significant at 10%; * significant at 5%; ** significant at 1% Regression uses standard errors adjusted for company-level heteroskedasticy. Table 3 gives regression results. The posive significant coefficient on R&D indicates that the greater the level of R&D spending per patent the greater the international information flows. In essence, a big research project pulls in more information from overseas sources. The average firm s R&D elasticy is 0.204, indicating posive, diminishing returns to R&D spending 9 from 9 Calculated by the R&D coefficient, plus the R&D*PRODUCTION coefficient times the mean of the PRODUCTION term. 7

8 knowledge spillovers. The negative significant coefficient on R&D*PRODUCTION indicates that as a firm produces more, the returns to R&D (from information flows) decreases. The posive significant coefficient on PATENTS shows that the more patents a firm gets, the more information uses from other countries; the posive significant coefficient on PRODUCTION shows that the more a firm produces abroad (relative to overall patenting), the greater the information flows from that country. Chrysler is not significantly different from General Motors as the coefficient is on CHRYSLER is not significant and GM is the dropped dummy variable. However, the posive and significant coefficient on FORD shows a statistical difference between Ford s and GM s information flows. This could be because Ford s smaller R&D program causes to rely more on using ideas from s acquisions. The intercept term and the GERMANY dummy show that flows from Germany start wh a disadvantage but increased production there leads to greater information flows compared to the same level of production from Canada/U.K. (as seen by the PRODUCTION*GERMANY coefficient). This yields an elasticy of addional production per patent on information flows from Germany 10 of 0.138, compared to on information flows from Canada/U.K. V. CONCLUSION The above results show that international M&As do indeed facilate information sharing, as a greater amount of manufacturing exposure in any of the countries leads to larger utilized knowledge spillovers to the home company. Thus, M&As by U.S. auto manufacturers cause them to use inventions from the countries of their acquisions more intensively than before in their own innovation process. However, the size of the impact on the U.S. firms innovation differs if the acquisions are in Germany rather than Canada/U.K. M&As in Canada/U.K. start wh an advantage in knowledge flows to the U.S. compared to those in Germany, perhaps due to the shared language. This fs wh Jaffe and Trajtenberg (1999) who find that information flows more easily between countries wh a common language. Language commonaly is only one determinant of the flow of information between countries. The relative advantage that Canada/U.K. begins wh erodes over time wh a firm s increased exposure to Germany, so that if a U.S. auto 10 Calculated by the PRODUCTION coefficient, plus the R&D*PRODUCTION coefficient times the mean of the R&D term, plus the PRODUCTION*GERMANY coefficient. 8

9 manufacturer were to increase s production presence in Germany through M&As, the resulting benefs to s own innovation process will be greater than if the manufacturer undertook a similar expansion in Canada/U.K.. This implies that having more M&As in another country might have a greater long term impact on a U.S. firm s innovation than a shared language. For this study, business involvement in Germany, wh s relatively large stock of technical knowledge in automobile production and engineering, eventually offsets any language difficulties for a U.S. auto manufacturer. Though enhanced information flows might not be a primary motivation in the international M&A decision, does show that such corporate restructuring can have a significant impact on the international information available for a manufacturer s innovative capabilies and that the location of the target firm exerts a strong influence on the size of the benef to the home firm s innovation. Isely and Simons (2002) find that the more U.S. auto manufacturers ce patents from Germany, the more patents the U.S. company obtains. Taken together wh the above results, this indicates that international M&As might lead to a greater overall volume of innovation by the acquiring firm. This research adds to the debate over issues that U.S. auto manufacturers face today. For example, in 2006 General Motors considered a merger wh Renault-Nissan. Would be beneficial for GM or s U.S. competors to undertake such M&As in the future? The above findings indicate that if a company s goal is to enhance s innovative capabilies then can do so by increasing s manufacturing exposure by merging wh or acquiring companies in other countries regardless of any language barrier. Whether this improved information flow leads to an increase in the firm s bottom line and future global competiveness is a question for future research. REFERENCES Belderbos, R. (2001) Overseas Innovations by Japanese Firms: An Analysis of Patent and Subsidiary Data, Research Policy, 30, pg Bessen, J. and Hunt R. (2004) An Empirical Look at Software Patents, Federal Reserve Bank of Philadelphia Working Paper No /R.. Danzon, P., Epstein A., and Nicholson S. (forthcoming) Mergers and Acquisions in the Pharmaceutical and Biotech Industries, Managerial and Decision Economics. Geroski, P., Machin S., and Van Reenen J. (1993) The Profabily of Innovating Firms, Rand Journal of Economics, 24(2): Ghosh, A. (2001) Does Operating Performance Really Improve Following 9

10 Corporate Acquisions, Journal of Corporate Finance, 7(2): Ghosh, A. (2004) Increasing Market Share as a Rationale for Corporate Acquisions, Journal of Business Finance & Accounting, 31(1/2): Giedeman, D., Isley P., and Simons G. (2006) Innovation and the Business Cycle: A Comparison of the U.S. Semiconductor and Automobile Industries, International Advances in Economics Research, 12(2): Granstrand, O. and Sjölander S. (1990) The Acquision of Technology and Small Firms by Large Firms, Journal of Economic Behavior and Organization, 13(3): Hall, B., Jaffe A., and Trajtenberg M. (2001) The NBER Patent Cation Data File: Lessons, Insights and Methodological Tools, NBER Working Paper Hall, B. and Ziedonis R. (2001) The Patent Paradox Revised: An Empirical Study of Patenting in the U.S. Semiconductor Industry, , , Rand Journal of Economics, 32: Isely, P. and Simons G. (2002) Global Influences on U.S. Auto Innovation, Economics of Innovation and New Technology, 11(1): Jaffe, A. (1986) Technological Opportuny and Spillovers of R&D: Evidence from Firms Patents, Profs, and Market Value, American Economic Review, 76(5): Jaffe, A., Fogarty M., and Banks B. (1998) Evidence from Patents and Patent Cations on the Impact of NASA and Other Federal Labs on Commercial Innovation, Journal of Industrial Economics, 46(2): Jaffe, A. and Trajtenberg M. (1999) International Knowledge Flows: Evidence from Patent Cations, Economics of Innovation and New Technology, 8: Jaffe, A., Trajtenberg M., and Fogarty M. (2000) Knowledge Spillovers and Patent Cations: Evidence from a Survey of Inventors, American Economic Review Papers and Proceedings, 90(2): Jaffe, A., Trajtenberg M., and Henderson R. (1993) Geographic Localization of Knowledge Spillovers as Evidenced by Patent Cations, Quarterly Journal of Economics, 108(3): Pavt, K. and Soete L. (1997) International Differences in Economic Growth and the International Location of Innovation, in Wolff, E. (ed.), The Economics of Productivy, Vol. 1. Cheltenham, UK: Edward Elgar Publishing. 10

11 Sundarsanam, S. (2004) Discussion of Increasing Market Share as a Rationale for Corporate Acquisions, Journal of Business Finance & Accounting, 31(1/2): Trajtenberg, M., Henderson, R., and Jaffe A. (1997) Universy Versus Corporate Patents: A Window on the Basicness of Invention, Economics of Innovation and New Technology, 5(1): Van Beers, C. and Sadowski B. (2002) The Impact of Acquisions and Divestures on Acquiring Firms Innovation, Working paper, Delft Universy of Technology, Department of Economics of Innovation. Ward s Automotive Yearbook ( ) Ward s Communications, Detro, MI. 11